Estimation of Victimization Prevalence Using Data from the National Crime Survey

1. Introduction.- 1.1. Cheery Indicators.- 1.2. Aims of the NCS.- 1.3. The Sample Design and Rotation Scheme.- 1.4. The Interview.- 1.5. Data Tapes.- 1.6. Missing Data.- 1.6.1. Types of Nonresponse.- 1.6.2. The Magnitude of the Nonresponse.- 1.6.3. Ignorable Sampling and Nonresponse Mechanisms.- 2. Some Intuitive Prevalence Estimators.- 2.1. Two Ad Hoc Estimators.- 2.2. The BJS Estimators.- 3. Consistency of the Ad Hoc Models.- 3.1. Modelling Victimization.- 3.2. Consistency of the Ad Hoc Estimator.- 3.3. The Modified Version of the Ad Hoc Estimator.- 3.4. Consistency of the BJS Estimators.- 4. Model-Based Estimators.- 4.1. A Homogeneous Bernoulli Model.- 4.2. A Correlated Bernoulli Model.- 4.3. A Markov Model.- 4.4. Comparisons of Estimates.- 5. Nonresponse.- 5.1. Checking the Missing at Random Assumption.- 5.2. Taking the Nonresponse into Account.- 6. Further Research and Conclusions.- 6.1. Extensions.- 6.1.1. Estimation at the HH Level.- 6.1.2. Possibilities for Further Research.- 6.2. Conclusions.- Appendix A. Questionnaire.- Appendix B. Computer Programs.